Update app.py

app.py

@@ -22,7 +22,7 @@ metadata = pickle.load(open("metadata.pkl", "rb"))
 # Load Hugging Face LLM (CPU-friendly)
 # -----------------------------
 # Small model for HF Spaces CPU limits
-model_name = "TheBloke/vicuna-7B-1.1-HF"  # You can replace with a smaller model if needed
+model_name = "TheBloke/vicuna-3B-1.1-HF"  # You can replace with a smaller model if needed
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")  # Hugging Face will manage CPU/GPU
 generator = pipeline("text-generation", model=model, tokenizer=tokenizer, max_length=200)